Computer vision guided laser targeting

The goal of this project is to use computer vision to autonomously aim a laser pointer at a specific point visible in a camera image. The target application is non-invasive measurement of human vital signs (heart rate, breathing rate, etc) using a doppler laser vibrometer. The approach taken is to solve for the mathematic relationship uniquely mapping every (x,y,z) point in a real world coordinate system to the unique controls (pitch, yaw) which aim the laser at this point. Two approaches are discussed based on usage of a (a) 3d depth sensor or (b) dual camera setup.

Downloads


Thesis Paper

Thesis Presentation

laser_overview.png

laser_mirror.png